Asiana 777 crashes in San Francisco

[Phil Bunch]

A video is available here:

http://www.cnn.com/2013/07/06/us/california-plane-incident/index.html?hpt=hp_t1

Basic news story here:

http://www.guardian.co.uk/world/2013/jul/06/san-francisco-airport-plane-crash-boeing

At the moment (623pm, KJFK time zone), there isn't much info about injuries, but apparently there are at least 2 dead and 70 injured, per CNN TV news.  A large fireball was reported and the airliner looks pretty badly burned from a helicopter TV video stream - large holes in the roof with a burned out interior visible.  323 were reported as souls on board by one news story.

Allegedly, no distress calls by the crew were made.  One pilot on the TV news said the plane landed short, but the TV video feed was too close-up for me to see that.  It is sitting on the ground next to the runway on TV news, with its tail assembly completely missing and not visible in the fairly close-up TV video stream.  Runway 28L is allegedly their intended runway.  They allegedly clipped the seawall on the way to a landing.  Many pieces of the tail assembly are visible near the end of the runway.  A pilot on TV speculates that the pilot tried to pull up once he realized they were too low and speculates that this caused the tail to impact near the end of the runway where many tail assembly parts including both horizontal stabilizers are now visible.

Still photos on TV show a number of people walking away from the crash site, so apparently many or most were able to evacuate in an emergency procedure - the inflatable slides are visible.

[Jeroen Hoppenbrouwers]

As always, better coverage here:

www.avherald.com

[Phil Bunch]

An infographic by the NY Times is linked below.  Not sure that it adds much info:

http://www.nytimes.com/interactive/2013/07/06/us/where-asiana-flight-214-came-to-rest.html?smid=fb-nytimes&WT.z_sma=US_WAF_20130706&_r=1&

The associated news article is here, with some additional information and comments:

http://www.nytimes.com/2013/07/07/us/san-francisco-plane-crash.html

Here's an excerpt, discussing the displaced threshold, and speculation that the captain may have forgotten about this detail if he used a visual approach, etc:

"Arnold Reiner, a retired airline captain and the former director of flight safety at Pan Am, said that it appeared from television images that the jetliner had touched down far earlier than the normal landing point, which is about 1,000 feet down the runway. That runway, 28 Left, has a "displaced threshold," he said, meaning that the runway's usable area does not begin at the start of the pavement. The Instrument Landing System would normally guide the pilot to the proper touchdown point, but in clear weather, pilots will sometimes fly a visual approach. "

[Jeroen Hoppenbrouwers]

ILS is inop due to upgrade. Everybody flies visual. The displaced threshold isn't very long at 28L and even then you rather clearly see where the water ends and the concrete begins. This wasn't a simple misjudgement, I fear.


Hoppie

[torrence]

Hypothetical, Jeoren,

What would happen if you didn't remember the ILS INOP and programed an autoland with incorrect threshold? What alerts/alarms would go off first?

Another random thing that occurred to me was what if one entered a totally bad Vref and the throttles followed that cue and stalled you early.  Any warning before the stick shaker? I know there are multiple checks to prevent a low and slow approach, but it appears that's what may have happened.  

Seems clear they were below glide path, stalled and dropped way early, but WHY?

Thank God the evac and other stuff worked as well as it did.

Torrence

Edit:  Agree about the visual - even as pax, on a clear day it's hard to miss the fact you're touching down just after crossing to dry land!
T

[Jeroen Hoppenbrouwers]

(speculation)

It will totally depend on whether the ILS glideslope had been turned off, or had been misaligned. In principle a glideslope beacon turned off would have prevented the aircraft from locking on and should have kept it at the 2000 ft GS intercept height. If however the GS beacon had been misaligned, it could potentially have been set up at an approach angle lower than 3 degrees which could have caused a lower than usual approach and potentially the crash as it happened.

However this is not very likely. The seawall was impacted off centerline, suggesting that the pilots flew manual. But this misalignment may have been caused by last minute (second) panic and pullup without thrust and other fumbling. If indeed the aircraft had been following a, say, two degree GS (I did not do the math), the indications would have been a way-off GS intercept (far further out than usual, many miles), higher pitch angle, lower vertical speed, and higher engine thrust on approach. The radar altimeter would have directed the autothrottles to go idle at the usual 20-30 ft over the sea, which now means way before the seawall.

It would be a 'nice' experiment to set the GS to 2 degrees and run the sim and see what happens.

We will have to be patient.

But ... an operative GS combined with a GS misalign and a tired crew still flying the ILS may explain things. It goes sort of okay until the A/T pulls the power off before the seawall and then you are too low to do anything. And tower people may not warn you as they expect you are flying visual and 'anything goes'. Don't forget that a two-degree glideslope actually will bring the ILS antenna (in the 744 these are on the nose gear door for finals; 777 may have the same setup) correctly to the touchdown point. It is the lower-hanging parts of the aircraft that hit ground before the touchdown point. Pilots are high up. There definitely is a different view, you are 1/3 lower than usual, but if you are tired...

It's a rainy Sunday. I may actually do the math. If somebody else starts it earlier, please post here.


Hoppie
(end of speculation)

[Jeroen Hoppenbrouwers]

Somebody plotted the FlightAware data against Google Earth.

https://twitter.com/sbaker/status/353611787750494208/photo/1

Although this isn't officially confirmed data, it does suggest a higher and steeper than usual approach if anything. Definitely not a low approach angle.

Near-flat segment on short final, followed by steep descent into the seawall. Could mean anything.


Hoppie




Picking up the replay boxes probably was easy this time...

[Jeroen Hoppenbrouwers]

Apparently the NTSB guys have already played back most of the stuff (which should be easy given the pristine condition of the boxes).

From AVHerald.com:

On Jul 7th the NTSB reported in a press conference at San Francisco Airport, the crew was cleared for a visual approach to runway 28L, the crew acknowledged, flaps were set at 30 degrees, gear was down, Vapp was 137 knots, a normal approach commenced, no anomalies or concerns were raised within the cockpit, 7 seconds prior to impact a crew member called for speed, 4 seconds prior to impact the stick shaker activated, a call to go-around happened 1.5 seconds prior to impact, this data based on a first read out of the cockpit voice recorder. According to flight data recorder the throttles were at idle, the speed significantly decayed below target of 137 knots - the exact value not yet determined -, the thrust levers were advanced and the engines appeared to respond normally. The NTSB confirmed the PAPIs runway 28L were available to the approaching aircraft before the accident, however were damaged in the accident and thus went out of service again. The localizer was available, the glideslope was out of service, according NOTAMs were in effect. There were no reports of windshear and no adverse weather conditions. The air traffic controller was operating normal, no anomaly was effective, until the controller noticed the aircraft had hit the sea wall. The controller declared emergency for the aircraft and initiated emergency response. ARAIB and Asiana personnel have arrived on scene and have joined the investigation. The Mayor of San Francisco reported runway 10L/28R was cleared for service.

Given the observed but not yet confirmed steep approach, I sincerely hope this wasn't a simple "oops" where the crew forgot to add thrust once the PAPI glide slope was captured.


Hoppie

[Jeroen Hoppenbrouwers]

There nowadays always is somebody filming anything.

http://www.youtube.com/watch?v=sEDZerwU7uE

[G-CIVA]

NTSB press conference:

http://youtu.be/XLYeUbeyfOg

NTSB Twitter page (for more info & pictures):

https://twitter.com/NTSB

[Phil Bunch]

The Wall Street Journal is reporting what I would call "apparent pilot error", informally:

http://online.wsj.com/article/SB10001424127887323823004578592152181919438.html

Here's an excerpt:
--------------------------------------------------

Pilots of the Asiana Airlines jet that crashed at San Francisco International Airport over the weekend allowed the Boeing 777's speed to dip dangerously low, and then apparently ran out of time to correct their landing approach, according to preliminary data released by investigators.

While stopping short of pinpointing pilot error as the likely cause of the fiery crash that killed two teenage passengers and injured dozens of others, National Transportation Safety Board Chairman Deborah Hersman indicated Sunday that investigators already are focused primarily on understanding why the crew allowed speed to decay to such an extent—and failed to take decisive action until the wide-body jet was less than two seconds from impact.

In the first on-scene briefing by the NTSB, Ms. Hersman said a preliminary readout of the plane's flight-data and cockpit-voice recorders indicates that everything was normal—with no discussion of any onboard problems or concerns about the visual approach in good weather—until just seven seconds before impact.

At that point, she told reporters, the crew realized the plane arriving on an overnight flight from Seoul was flying too slowly. "The speed was significantly below" the designated approach speed of roughly 130 miles an hour.

Three seconds later, a stall-warning activated, indicating the Boeing 777 carrying 307 passengers was losing aerodynamic lift. The crew didn't act to sharply increase engine thrust and try to climb away from the strip—conducting what is called a "go-around"—until 1.5 seconds before impact.

The safety board believes "the engines appear to respond normally" to those commands, Ms. Hersman said, but by then it was too late to recover, and portions of the lumbering jet slammed into the seawall in front of the strip.

Taken together, the preliminary data and Ms. Hersman's early description of the sequence of events strongly suggest investigators are leaning away from mechanical or other system failures as the likely culprit.

During the media briefing, Ms. Hersman stressed that there was "no prior distress call" from the cockpit crew, which should have been able to rely on both ground-based and onboard landing aids to "establish an approach path" to safely reach the beginning of the runway. The jet slammed down roughly 1,000 feet short of that point, leaving some parts on the seawall separating the runway from San Francisco Bay. Other parts were recovered from the nearby water.

Because the jet was on a visual approach in excellent weather, "you don't need instruments to get into the airport" safely, Ms. Hersman said.

In another significant disclosure, she said investigators "haven't identified any specific similarities" between Saturday's Asiana crash and the 2008 belly landing of another Boeing 777, operated by British Airways, on final approach to London's Heathrow International Airport. Investigators determined the British jet's engines were starved of fuel after chunks of ice blocked its fuel system.

[farrokh747]

In the ATC recording released, you can hear radio comm between the acft and tower for some time after impact - so the comm system (i guess VHF) was still working after the crash? Perhaps power switched to the batt buses automatically?

[Jeroen D]

Given the observed but not yet confirmed steep approach, I sincerely hope this wasn't a simple "oops" where the crew forgot to add thrust once the PAPI glide slope was captured.

Early days still, but it would not be the first time.
Jeroen

[Jeroen Hoppenbrouwers]

10,000 hour seasoned captain flying, including 747-400 experience, but still training on the new 777 (40 hours under his belt). Supervised by a 3000+ hour 777 training captain. Rats. This should have been about the safest landing ever.

I fear the training department will get a few remarks very soon.


Hoppie

[martin]

On PPruNe, someone described a tricky feature of the 777 for the following case:

¤ autopilot (AP) is ON, but not in a "speed" mode
__(FLCH was mentioned in particular, and described as a "trap" in this context);
¤  autothrottle (AT) is ON.
¤ Now AP is disconnected, but
¤ AT is left ON (which seems to be customary on the 777),

In this case (if I understood correctly), the AT while "technically" still ON will actually do nothing any more; you'll have to move the throttle levers manually to change thrust.

Not a problem as long as you are aware of this situation, but if not...
(Note that no one is suggesting this actually played a role in the Asiana crash.)

Is this behaviour the same on the 744?

Martin

[farrokh747]

Q: is there a difference in disconnecting the AT from the TQ switch, and disarming the AT from the MCP switches (L&R for the 777)

So lets say 137kts is set in the MCP, AT switches on MCP are armed, and AT is disengaged via the TQ switch, will the AT kick in at any point as the speed trend approaches 137kts, assuming the levers are at idle, without any intervention?

fc

[John Golin]

(My understanding is no - it will not 'reengage' .  Interestingly VNAV does have some altitude protection with AT disengaged - http://www.youtube.com/watch?v=IJKU19Rn0GA - not what I expected until I saw the clip ).


I've never understood why the AT, by design, can be 'on' and not active in a mode - and why HOLD exists (outside of takeoff for both).

[Joe Clark]

(My understanding is no - it will not 'reengage' .  Interestingly VNAV does have some altitude protection with AT disengaged - http://www.youtube.com/watch?v=IJKU19Rn0GA - not what I expected until I saw the clip ).


I've never understood why the AT, by design, can be 'on' and not active in a mode - and why HOLD exists (outside of takeoff for both).

Is the AT "on" or is it simply "armed"? To me, when the L/R AT switch is pushed up, the AT is "armed"  When the A/T button is pushed and the light is on, the AT is on.

[Phil Bunch]

Here's a link to a press article on the "FLCH trap", as it's being called here and there.  They make reference to pprune.

http://www.theatlantic.com/national/print/2013/07/professional-pilots-on-the-san-francisco-crash/277563/

PAPI being active until apparently wrecked by the crash itself is yet another factor that makes all of this seem so incomprehensible to this lay person.  

And we thought the Air France high-altitude stall over the Atlantic was an anomaly...in some ways, this seems worse to me.  It's hard to beat setting the throttles to idle and apparently flying the thing into the ground on a clear day.  The stick shaker alarm appears to be a bit late for this specific situation.  

We can't remind ourselves too many times to wait for the final investigation report, but it's also very hard to avoid some degree of informal speculation.
-----------------------------------------------------------

How can the aviation community possibly prevent these very rare events where the crew has some sort of group delusional experience and screws up what seems in retrospect to be an "obvious" problem?  Perhaps we are at an accident rate that is as low as it can practically be?  We don't know how to write perfect software either, so more automation isn't the answer, I personally believe.  Would better training really fix these rare, seemingly inexcusable problems?

[John H Watson]

So lets say 137kts is set in the MCP, AT switches on MCP are armed, and AT is disengaged via the TQ switch, will the AT kick in at any point as the speed trend approaches 137kts, assuming the levers are at idle, without any intervention?

From the 777 Maintenance Manual

Automatic Speed Mode Engagement

If the A/T is not engaged, it will automatically engage in the speed (SPD) mode when all these conditions are true:

*At least one A/T ARM switch is in the ARM position
*Radio altitude is more than 100 feet, or more than 400 feet after takeoff
*Engine thrust is below the engine thrust limit
*The autopilot or the flight director are not in a speed-through-elevator mode
*The airspeed is just above stall speed

Without knowing what mode the FD was in, it would be hard to say if the A/T would have engaged.